Ceph » rgw

When trying to set a new bucket lifecycle using the AWS SDK, the request fails with "501 Not Implemented".

PUT /pdtest_expire_test?lifecycle HTTP/1.1
[…]

<?xml version="1.0" encoding="UTF-8"?>
<LifecycleConfiguration xmlns="http://s3.amazonaws.com/doc/2006-03-01/"><Rule><ID>test</ID><Status>Enabled</Status><Expiration><Days>1</Days></Expiration><Prefix></Prefix></Rule></LifecycleConfiguration>

HTTP/1.1 501 Not Implemented
[…]

<?xml version="1.0" encoding="UTF-8"?><Error><Code>NotImplemented</Code><RequestId>tx000000000000000001e40-00589cff17-6f1d60-default</RequestId><HostId>6f1d60-default-default</HostId></Error>

The cluster is running version 11.2.0.

After increasing the log level in one rgw instance I can see following lines:

2017-02-10 01:08:51.225783 7fdc60167700 10 delaying v4 auth
2017-02-10 01:08:51.225785 7fdc60167700 10 ERROR: AWS4 completion for this operation NOT IMPLEMENTED
2017-02-10 01:08:51.225788 7fdc60167700 10 failed to authorize request
2017-02-10 01:08:51.225789 7fdc60167700 20 handler->ERRORHANDLER: err_no=-2201 new_err_no=-2201

Using s3cmd with V2 signatures, the request is successful. Forcing s3cmd to use V4 signatures, the request fails again.

As support for V2 signatures is fading out and therefore gets removed from AWS SDK I think being able to use V4 signatures with radosgw for as much operations as possible would be a good thing.